Hi Ben,
Seems to me that these values are calculated since boot.
So what you have to do is record the value, then look at it a bit later,
and use the difference to calculate your values.
Hope this helps,
Arend
On Tue, 25 Sep 2001, Ben Logan wrote:
> Hello,
>
> I would like to write a simple cpu monitoring program and need to get
> the current cpu stats.
>
> It looks like I could get the info from /proc/stat, but I'm having
> trouble makeing sense of it. I thought I had figured it out, but the
> numbers I get don't change with system load. As an example, here's
> the first line of my /proc/stat file:
>
> cpu 278485 11148 67657 4231952
>
> According to the proc(5) manpage, these numbers are "The number of
> jiffies (1/100ths of a second) that the system spent in user mode,
> user mode with low priority (nice), system mode, and the idle task,
> respectively." I thought that by adding fields 2-5, multiplying by
> 100 and dividing field 5 by the result, for example, would give me the
> current percentage of idle cpu time. It comes pretty close to top's
> statistic, but doesn't match exactly and doesn't go down when I load
> the system.
>
> Then I thought maybe I shouldn't include the second field as it seems
> it would be a subset of field 1, but that didn't help.
>
> Here's the command I ran to test things:
>
> gawk '$1=="cpu" {total=$2+$3+$4+$5; print("Idle: "$5*100/total)}'
> /proc/stat
>
> I'm running Seawolf with kernel 2.4.9-ac14 on a single-processor
> Intel box.
>
> Anyone know of any good sources of info on this? (I don't know C, so
> the source would probably not help me much. :)
>
> Thanks,
> Ben Logan
>
_______________________________________________
Seawolf-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/seawolf-list
